Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/78446
Title: Sound source localization in highly reverberant environment based on sparse Bayesian framework
Authors: Ge, Yihui
Keywords: DRNTU::Engineering::Electrical and electronic engineering::Electronic systems::Signal processing
Issue Date: 2019
Abstract: Sound source localization is one of key techniques in audio signal processing, while achieving indoor localization is known to be challenging due to its high computational complexity. In this paper, a localization algorithm based on sparse Bayesian framework is proposed to solve the problem of localization in reverberate environments. The algorithm achieves localization by dividing the detected area into grids and constructing a parametric dictionary with parameters being unknown reflective ratios of the enclosure. After that, methods of variational Bayesian Inference are used to approximate to the actual value of unknown parameters. The proposed algorithm has advantages that it proposes a multi-parameter model with parameters being reflective ratios of walls, which is critical, but generally unknown in actual localization. Besides, parametric tuning steps are replaced by statistical methods to improve efficiency and accuracy. During the numerical simulation, the algorithm is proved to have the property of rapidity and accuracy.
URI: http://hdl.handle.net/10356/78446
Schools: School of Electrical and Electronic Engineering 
Research Centres: Centre for Infocomm Technology (INFINITUS) 
Rights: Nanyang Technological University
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:EEE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
FYP+Final+Report+Ge+Yihui.pdf
  Restricted Access
1.38 MBAdobe PDFView/Open

Page view(s)

276
Updated on Jun 14, 2024

Download(s)

5
Updated on Jun 14, 2024

Google ScholarTM

Check

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.